Graph Analytics in Snowflake: Beyond Traditional SQL
summary
Neo4j Graph Analytics on the Snowflake Marketplace runs graph algorithms directly against Snowflake tables with a zero-ETL approach, doing graph projection, algorithm execution and result-writeback in a single SQL call. It requires node ID and source/target columns (ints or strings), numeric property columns, and a single label per node. A demo runs PageRank on flight data, spins an XS compute (one core, 6 GB) via the compute pool selector, writes scores back to Snowflake and visualizes them — the job completes successfully and an Estimation API returns JSON pool recommendations.
AI-written summary · about
chapters (15)
- 0:00Introduction and Speaker Welcome
- 0:57Understanding Snowflake: The Data Cloud
- 4:38Important Disclaimer
- 7:01Anatomy of an Algorithm Call (The 3-Step Process)
- 8:37Compute pool selection
- 9:53Modeling Tabular Data as Graphs
- 11:47Schema Requirements for Node and Relationship Tables
- 13:17Table of differences
- 14:55Overview of Available Graph Algorithms
- 15:24Result Types Overview
- 16:55Estimation API
- 18:05Visualisation Capabilities
- 18:45Live Demo
- 22:13The Job Life Cycle and Architecture
- 23:45Future Integrations: Snowflake Cortex AI Agent. Upd.: Now available for preview
related talks
27mNeo4j GDS for Snowflake
23mAcceleratinag Data Science and Machine Learning with Snowflake DataTech
13mLightning Talk: Highlights of the Snowflake Elastic Data Warehouse
36mImplementing Natural Language Analytics with Snowflake Cortex
18mGraph analytics: From zero to hero
01h 13mBuilding a Semantic AI Layer in Snowflake: From Metrics to AI Insights
29mAll about graphs in Database
21mCaching Snowflake data with Aerospike for advanced analytics